Power of One

TTC ran 40+ brand websites as disconnected silos. Conversion rates varied by 300%+ across the portfolio. Getting a feature shipped took 18 months. Every brand duplicated the same work.

The outcome: a single repeatable B2C ecommerce engine that reduced migration time to 6 weeks, improved baseline conversion by 12% across pilot brands, and cut design and engineering duplication costs by £2M annually. The same platform now carries 2M+ travellers a year.

The Travel Corporation is a $2bn group operating 42 travel brands across 60+ countries — including Trafalgar and Contiki — carrying over 2 million travellers annually.

In late 2019, 80–90% of bookings happened through agents or by phone. The websites were expensive brochures, not conversion tools. Each of the 40+ brands had built its own site independently, accumulating 10+ years of duplicated decisions, inconsistent patterns, and fragmented user experiences.

When COVID halted travel entirely in early 2020, it created a window. Customers were spending more time online, building confidence with digital platforms. If TTC could shift to a genuine B2C model before the recovery, it would be in a stronger position than competitors who waited.

The Power of One programme had been approved just before COVID hit. The data team and additional design resource were cut as a cost measure. I joined in February 2020 as the founding designer, and we pressed ahead.

Brand politics blocked the clean-sheet principle

The biggest risk wasn’t technical — it was organisational. Brand CEOs had defended accumulated decisions for over a decade. “We’ve always done it this way” was the default response to change.

I ran analytics audits across all 40+ brands and presented each brand’s conversion data alongside competitor benchmarks. That reframe shifted conversations from opinion to evidence. It was the hardest and most important sell of the project.

The business rejected the original design direction

Our first proposal was a progressive web app approach with a stepper-style booking flow. The business pushed back hard — it was too different from what brands knew.

We refactored to a more familiar structure while preserving the underlying performance and conversion improvements. The lesson: high-fidelity mockups unlocked approval. Brand stakeholders made visual decisions. Abstract architecture didn’t land.

One system for 40+ identities

The design system needed to serve brands as visually distinct as a luxury escorted touring company and a youth adventure brand. Too rigid and brand teams would reject it. Too flexible and it would fragment within months.

I separated structure from style using a token architecture. Core components, spacing, and interaction patterns were shared. Colour, typography, and visual identity lived in brand-specific theme layers.

Lead Product Designer. Founding designer on the central PO1 task force, spanning Product, Engineering, Data, CRM, Marketing, and Operations.

What I owned:

• UX architecture across the full booking journey
• Design system built from scratch — Figma, Storybook, Zeroheight documentation
• The five design principles that governed every decision for the life of the programme
• Stakeholder alignment across 40+ brand CEOs and the executive steering group
• The analytics audit that secured executive approval for the clean-sheet approach
• Weekly design-data-engineering sync, established six weeks in when parallel workstreams began to diverge

What I contributed to:

• Discovery research sprint (15+ user interviews, competitor benchmarking)
• Scope governance — the framework that blocked feature creep throughout delivery
• Brand migration process design at each stage

Why a Principal-level designer was needed here

40+ brand stakeholders defended 10+ years of legacy decisions. The role required someone who could challenge opinion with evidence, design at a systems level, and work as a strategic partner to product and engineering leadership. A senior IC designing screens wouldn’t have moved the organisational blockers.

We ran an intensive research sprint at the start of the project: 15+ user interviews, customer data analysis across brands, competitor benchmarking, and a SWOT analysis of the current market offering.

Key findings:

• 80–90% of guests still booked through agents or by phone, not online
• Average guest age was 63, with a wide range of digital confidence
• Group size, social dynamics, and travel confidence were primary purchase drivers
• Navigation and search were the biggest friction points on existing sites
• Price clarity and booking confidence were the most common drop-off causes

Analysing brands in isolation had hidden three critical drop-off points that only became visible when we looked at the portfolio as a whole.

These weren’t values on a wall. They were decision filters applied at every fork in the road:

1. Start from a clean sheet — no legacy decisions inherited by default
2. Design for B2C only — no hybrid compromises that serve neither audience well
3. Prioritise performance and conversion — speed and outcome above aesthetic preference
4. Validate assumptions before building — test before committing engineering time
5. Build once, reuse everywhere — the system pays for itself in every migration

Every scope request got tested against these. “Does this serve B2C conversion?” became the question that ended most additions quickly.

Progressive web app vs. familiar web patterns

Our first direction was a PWA with a stepper booking flow. The business rejected it — too different from what brands knew. We refactored to familiar patterns while keeping the performance improvements underneath. The trade-off: a more conservative visual approach in exchange for faster approval and delivery.

Design system first vs. design system later

On a tight deadline, building the system upfront felt like a risk. We did it anyway. Clean handoff to engineering required it, and retrofitting components later would have cost more time than it saved.

Mobile-first vs. desktop-first

Research showed users started their journey on mobile but often completed bookings on desktop. We designed mobile-first to ensure that entry point was solid, then scaled to desktop. This shaped information hierarchy and scroll depth across every template.

The core design system challenge was serving 40+ visually distinct brand identities without building 40+ different systems.

I separated structure from style across three layers:

• Core tokens — spacing, type scale, grid, interactive states
• Semantic tokens — colour roles, component-level decisions
• Brand theme layer — Trafalgar, Contiki, and every other brand applying their identity through the top layer alone

One change at the core propagates across every brand. One brand can update its visual identity without touching the underlying system.

The component library was built in Figma with Storybook integration and documented in Zeroheight. Handoff standards and QA processes were built alongside the components — not added later.

Business

Metric	Before	After
Brand migration time	18 months	6 weeks
Baseline conversion	Baseline	+12% across pilots
Design/engineering duplication	High	£2M annual saving
Page load performance	—	+40% improvement
Design-to-dev handoff time	—	−70%

User

Metric	Result
Accessibility compliance	30% → 95% WCAG 2.1 AA
Bounce rate	−15%
Average time on tour pages	3min 47sec (excl. bounces)
RAQ conversion on tour pages	0.22%*

*£5k+ guided tours don’t convert like retail. RAQ (request a quote) was the intended north star metric at MVP stage. The 3min 47sec average engagement time on tour pages indicates strong intent from the right audience.

Org and process

The first enterprise-wide design system supporting 2M+ guests annually. One design team capable of supporting 40+ brands. A repeatable discovery and validation process embedded across every brand migration. Formal scope governance that blocked feature creep throughout delivery.

Instrument analytics earlier. I instrumented analytics after launch, not before. Early data from pilot brands was incomplete as a result. Tracking requirements should be embedded in the design system from day one — not added in QA.

Build stakeholder coalitions before design begins. I underestimated brand politics in the first six weeks. Too much time in Figma, not enough time mapping stakeholders and pre-selling the clean-sheet principle. A structured stakeholder mapping exercise at the start would have saved weeks.

Push harder for a proper research budget. We lost the data team early due to COVID cost cuts. Guerrilla testing worked but couldn’t reach the primary demographic properly. Typical guests were 60+ year-olds booking £5k+ trips — hard to recruit cheaply. We should have pushed harder to protect the research budget when the cuts came.

At Principal level, your most important output isn’t the design. It’s the conditions that allow good design to happen repeatedly without you.

The governance model, the principles as decision filters, the scope document as a contract — these had more long-term impact than any individual component. That shaped how I approached every subsequent role.

The token architecture was built for visual consistency. The next step is instrumented components: design tokens tied to analytics so teams can measure component performance in production, not retrospectively.

With enough platform data, there is also a real opportunity to move from pattern reuse to personalised journey assembly. The infrastructure exists. The data layer needs investment.